If AI coding costs $750/month but replaces a $120K salary, who captures the savings?
GitHub Copilot’s first month of usage-based billing shocked developers with 10-50x cost increases — yet even at $750/month, the AI is cheaper than the junior developer it replaced. As AI tools price themselves based on actual value delivered rather than flat subscriptions, who benefits from the enormous gap between AI costs and human salaries: the companies, the remaining developers, or the displaced workers?
Commentaires (1)
Today’s Minds, Bodies, and Terawatts episode (July 1st) dug into the first real billing data from GitHub Copilot’s token-based pricing — and the math is striking. Even developers facing $750/month bills are paying a fraction of what a junior engineer costs, while 51% of committed code is already AI-generated. The episode explores what the book calls the ‘substitution threshold’: the moment when the tool costs so little relative to the worker that the economic argument for hiring becomes nearly impossible to make. Give it a listen and let us know — should the productivity gains from AI coding flow back to developers as higher wages, or are we watching the profession price itself out in real time?
Related reading on unscarcity.ai:
Envie d'aller plus loin ?
Obtenez le plan complet dans <em>L'ère de la post-pénurie : Repenser la société à l'ère des machines</em>
